Targeting lean process improvement projects for maximum financial
impact
John Darlingtona, Mark Francisb*, Pauline Founda and Andrew Thomasb
a
b
School of Business, University of Buckingham, Buckingham MK18 1EG, UK; Cardiff School of Management, Cardiff Metropolitan
University, Llandaff Campus, Cardiff CF5 2YB, UK
This empirical paper details a 12-month applied research project at a UK low-volume manufacturer of large vehicles. The
industry problem from which this study originates was a concern over the subjective nature with which the firm’s existing
lean intervention projects were being targeted (prioritised and selected). A structured literature review on this topic was
unable to identify any objective decision support mechanism for doing so; one that encompassed financial as well as
operational criteria. The resultant study was organised around an established seven-step action research framework. The
main body of evidence was derived from extensive analysis of financial and operational data extracted from the firm’s
enterprise resource planning system, along with two structured workshops that each involved multiple informants drawn
from the firm’s production centres and its accountancy department. Supplementary primary research was provided in the
guise of numerous unstructured interviews to validate data and from observation of shop floor practices. The main
contribution of this article is identifying and addressing the gap highlighted above, by developing and testing a financially
driven method for objectively targeting process improvement interventions within this large and geographically dispersed
operation. This innovative method includes five new constituent techniques.
Keywords: process improvement; lean; targeting; cost analysis; action research
1. Introduction and research context
This paper details a case study of a 12-month action
research (AR) project at a UK low-volume manufacturer
of large vehicles, during which an effective new lean
process improvement project targeting method was
designed, implemented and assessed. In the interest of
commercial confidentiality, this firm is referred to as
VehicleCo for the remainder of this paper. This case firm
is part of a large multi-national enterprise that is
headquartered outside of the UK. VehicleCo’s UK
operation is based upon six plants distributed throughout
the country. These design, manufacture and support a
variety of structures for the parent enterprise’s range of
products and also fabricate the component parts that
comprise these structures. At the time of the study reported
upon in this paper, this company had a portfolio of more
than 20 separate products/ size variants, which the firm
terms
‘contracts’. The structures for each contract are assembled
on dedicated production lines in its final assembly area. On
completion, these are shipped to another plant within the
parent group for the integration of various systems and the
kitting out of the vehicle interior.
VehicleCo had already amassed significant experience
of implementing lean (Womack, Jones, and Roos 1990;
Womack and Jones 1996) process improvement
intervention projects within its business. There was an
established and experienced internal lean process
improvement team composed of manufacturing engineers,
and they were working on a large ongoing project in
concert with a team from a global management consulting
firm. The focus of this project was waste reduction within
the final assembly area, particularly labour waste
reduction.
However, whilst significant waste and cost reduction
performance had been reported under the aegis of the firm’s
long-standing lean manufacturing programme throughout
its existence, there was growing concern that existing
techniques for targeting (prioritising and selecting) such
intervention projects within the firm were subjective in
nature. Two commonly heard metaphors that were aired in
conjunction with this problem were: ‘How do we know that
we are not picking the low hanging fruit’? – and – ‘Where
should we intervene next to assure the maximum process
improvement bang for our bucks’? This situation is clearly
undesirable as it risks sub-optimal management decisionmaking. Senior management were therefore seeking a
decision support method to target and hence evaluate the
impact of such projects within their lean programme; one
that was objective in nature and encompassed financial
criteria such as cost reduction potential.
Existing literature provides no such guidance. The aim
of this article is therefore to address this gap by developing
a suitable targeting method. Doing so would make a
valuable contribution to knowledge on the topical subject
of lean thinking. It would also offer a practical
contribution, it would greatly improve management
decision-making and the future economic impact of the
lean programme at VehicleCo. If the resultant method
proved to be generalisable, it would likewise offer
improved economic impact potential for all similar
manufacturing firms adopting this method to support their
indigenous lean initiatives.
The article starts by reviewing the literature on the
targeting and evaluation of process improvement
initiatives within the lean paradigm, including the evolving
nature of the kaizen (continuous improvement) concept,
and the role of value stream mapping (VSM) within this.
The second section details the research methodology that
was developed to realise the above research objective. This
includes a discussion of the AR strategy and the detailed
procedures that were used to collect, analyse and validate
the data. This is followed by a discussion of the findings
that were derived using this methodology. The paper
concludes with a summary of its academic and practical
contributions, methodological limitations and avenues for
future research.
2. Literature review – targeting lean process
improvement initiatives
2.1. The meaning of ‘Lean’ ‘Lean’ remains a topical subject
with four papers within this journal alone in 2014 (see
Lyonnet and Toscano 2014; Martinez-Jurado and MoyanoFuentes 2014; Vinodh, Kumar, and Vimal 2014; Wong,
Ignatius, and Soh 2014). A cursory bibliographic analysis
using its four main synonyms of lean manufacturing, lean
production, lean thinking and lean management reveals
that the literature on lean is composed of an extremely large
and growing body of material. It also reveals that a
significant proportion of this is composed of contributions
from practitioners and ‘gurus’, including many of its most
influential and highly cited texts (see Womack, Jones, and
Roos 1990; Womack and Jones 1996; Liker 2004; Dennis
2007).
The term lean itself was coined by the Massachusetts
Institute of Technology researcher John Krafcik whilst
working on the International Motor Vehicle Program, and
entered the management lexicon via his 1988 article in the
Sloan Management Review. Whilst coined by Krafcik,
Schonberger notes that many people attribute the origins of
the lean paradigm to the popular book by Womack, Jones,
and Roos (1990), although he asserts that lean production-
type initiatives were already well established in the US in
the early 1980s albeit under different names such as ‘JustIn-Time (JIT)’, ‘stockless production’ and ‘Zero Defects
(ZD)’ (2007, 406–408). Even though lean can therefore
boast a lineage of over three decades, it suffers from an
issue of interpretive viability (after Benders and van Veen
2001). Samuel (2011) suggests two related reasons for this
issue. The first is a lack of common definition within the
literature (Lewis 2000; New 2007; Shah and Ward 2007;
Bayou and De Korvin 2008). The second reason is that as
a concept, lean has evolved over time (Hines, Holweg, and
Rich 2004; Papadopoulou and Ozbayrak 2005). To these a
third reason might be added; a blurring of the boundaries
between the lean paradigm and similar contemporary
process-oriented operations paradigms such as the theory
of constraints (TOC), agility and six sigma – further
compounded by the emergence of hybrid paradigms such
as leagility and lean-sigma.
To overcome this interpretive viability dilemma,
authors have adopted numerous strategies to bound and
communicate their interpretation of lean. One approach,
such as that adopted by Dahlgaard and Dahlgaard-Park
(2006), is simply to utilise Womack, Jones, and Roos
(1990, 13) original characterisation:
Lean production is ‘lean’ because it uses less of everything
compared with mass production – half the human effort in
the factory, half the manufacturing space, half the
investments in tools, half the engineering hours to develop
new products in half the time. Also, it requires keeping far
less than half the needed inventory on site ….
A second strategy, adopted by authors such as Lewis
(2000) and Melton (2005), has been to invoke Womack
and Jones (1996) widely disseminated prescription of Five
Lean Principles for achieving leanness in lieu of definition.
Whilst a third, teleological approach, adopted by authors
such as Feld (2000) and Shah and Ward (2007) has been to
conceive leanness in terms of progress towards the
implementation of its constituent tools and techniques.
2.2. Evolution of kaizen within the lean paradigm
Regardless of the conception and definition of lean
(above), the concept of kaizen forms one of its integral
components. Kaizen is a Japanese term for continuous
improvement that was introduced by Imai (1986). It is the
umbrella concept for the execution of process
improvement within the lean paradigm, and is composed
of an underlying management philosophy and supporting
set of techniques and tools (Bicheno and Holweg 2008).
Kaizen is premised upon the continuous incremental
improvement of all work processes by all of the firm’s
employees (Imai 1986, 1997), with these latter two features
revealing its total quality control heritage. It is
implemented by worker teams that have often been
referred to as ‘quality circles’ (Karlsson and Ahlstrom
1996; Groover 2007). Such quality circles have been the
subject of widespread analysis and publication since the
early 1980s, when this technique first came to prominence
in the west and was often considered to be one of the
secrets of the success of Japanese manufacturing witnessed
in that period (see for example Hayes 1981; Schonberger
1982; Lawler and Mohrman 1985). However, whilst
quality circles were historically convened to address
quality and productivity problems, contemporary circles
have been used to address cost, safety, maintenance and
other concerns, with the term ‘Kaizen Circle’ (KC) often
being adopted to reflect this broader range (Dennis 2007;
Groover 2007). A KC typically comprises 6–12 volunteers
drawn from the same work area, who meet regularly to
respond to improvement suggestions or specific problems
encountered within that area. A lean programme will
therefore comprise multiple KC teams; all working to
resolve multiple concurrent problems. Each such KC is
scheduled to meet periodically in company time, with each
meeting lasting for at least an hour (Groover 2007).
Typically, team members are trained in the use of problem
solving tools such as the Seven Basic Quality Tools
(Ishikawa 1985) and 5 Whys, and to use these within
Deming’s (1986) Plan–Do–Check–Act problem solving
framework in order to systematically identify and redress
the root cause of the problem symptoms being considered
at that circle (Imai 1997; Bicheno and Holweg 2008).
In stark contrast to the incremental improvement
embodied by kaizen, the publication of Womack and
Jones’ seminal book on Lean Thinking in 1996 introduced
the term ‘kaikaku’, a Japanese word meaning ‘instant
revolution’ (Bicheno and Holweg 2008). The kaikaku
concept aims to rapidly realise radical process
improvement results, and is better known under a number
of synonyms such as ‘kaizen blitz’, ‘breakthrough kaizen’,
‘system kaizen’, ‘flow kaizen’ or ‘kaizen event’ (see
Womack and Jones 1996; Feld 2000; Melton 2005;
Bicheno and Holweg 2008). This has resulted in the
emergence of a lean process improvement typology, with
the traditional KC approach described above now often
labelled ‘process kaizen’ or ‘point kaizen’ to distinguish it
from this more radical and increasingly prevalent
improvement type. For example, Bicheno and Holweg
(2008) suggest that the process kaizen is concerned with
the elimination of localised waste and should be the
responsibility of shop floor workers. By contrast, they
suggest that the flow kaizen is concerned with the
stimulation of swift and even flow of throughput (after
Schmenner and Swink 1998) within the value stream, and
should be the responsibility of senior management. We
adopt this nomenclature for the remainder of our article.
2.3. Value stream mapping
To recap then, the process kaizen concept entails a
response to localised improvement suggestions or
problems by standing workplace quality circle or KC
teams. By contrast, the flow kaizen concept entails the
proactive design and implementation of discrete process
improvement ‘breakthrough’ projects, and therefore
suggests the need for temporary project teams and the
targeting of such interventions. This targeting imperative
is recognised by Nicholas (1998, 33), who states that ‘… an
organization must be able to target its improvement and
waste-elimination efforts. There must be some scheme for
putting priorities on where to expend time, effort and
resources [so that they contribute the greatest good]’. Put
simply, some objective means of deciding where to start a
lean programme within a firm, and once started, where to
intervene next where multiple options present themselves?
However, a structured literature review on this topic was
unable to identify any decision support method for doing
so.
Instead, this review highlighted the increasing
prevalence of flow kaizen projects and the appearance of
various VSM techniques since the latter part of the 1990s.
Indeed, VSM has emerged as the de facto vehicle for
implementing such projects and the most influential of
these techniques, as evidenced by both the extent of its
citation within the literature and implementation in
practice, is Rother and Shook’s (1998) Learning to See
approach (see for example Seth and Gupta 2005;
Abdulmalek and Rajgopal 2007; Serrano, Ochoa, and De
Castro 2008; Lasa, De Castro, and Laburu 2009;
Gurumurthy and Kodali 2011). By the same criteria, other
influential VSM techniques and publications have been
provided by Hines and Rich (1997), Tapping and Fabrizio
(2001), Jones and Womack (2002) and Duggan
(2002).
All of these VSM techniques share a number of
common features. For example, they are all premised upon
the actual mapping work being undertaken by a
multifunctional project team drawn from the areas of the
organisation believed to be transected by the value stream
concerned. Quite clearly, the conception of the value
stream to be mapped will have an important bearing on the
composition of the team selected. Likewise, all of the cited
VSM techniques stress the need for conceptual
simplification. This is expressed in the guise of the
mapping only a single, representative value stream rather
than attempting to capture the complexity of routing of the
organisation’s complete product portfolio. These
techniques also all share a discernibly similar project
anatomy. This includes a planning phase during which the
project team is formed, roles and responsibilities assigned
and the project’s terms of reference and focal value stream
established. This phase is usually followed by a workshop
to train the mapping team in general lean principles and the
specific mapping tool to be deployed. Next, comes the
‘current-state’ mapping phase during which the current
configuration and ‘baseline’ operational performance of the
focal value stream is established collectively by the team,
and described using a tool of prescriptive format and
standard iconography. This is followed by a ‘future-state’
mapping phase during which they collectively visualise its
idealised reconfiguration and consequential estimated
operational performance level, using the map tool itself as
a blueprint for this reconfiguration exercise. The
penultimate phase is again a collective enterprise, and
entails the team drawing up plans for a staged series of
intervention projects to realise the future-state blueprint.
The final phase is the implementation and monitoring of
these interventions.
Whilst VSM has become the de facto vehicle for
implementing flow kaizen projects, these techniques lack
guidance for conceiving or objectively selecting, from the
many alternatives that are typically available, the value
stream that is to act as the focus for the mapping exercise
itself. This selection will clearly have a profound bearing
both on the risks of successful implementation and the
potential performance improvement returns associated
with the investment in the consequent lean intervention
project. As indicated by the two metaphors cited during the
Introduction to this paper, it is notable that the lack of an
objective VSM targeting method has left many lean teams
in practice open to the accusation of sub-optimal decisionmaking, including the ‘cherry picking’ of easy to
implement interventions or the ‘pet projects’ of sponsoring
managers. Of all the VSM authors, only Tapping and
Fabrizio (2001, 13–15) address the issue of focal value
stream selection. They suggest three selection methods.
The first is subjective being merely to react to the most
vocal external customer demand for improvement. The
second is to undertake a product quantity analysis, which
involves producing a Pareto chart that illustrates the
distribution of production quantity by product. The authors
propose that any product representing more than 20% of
total production volume should be prioritised as a VSM
candidate. If the analysis of volume is inconclusive, or if
the variety of products and processes is complicated,
Tapping and Fabrizio (Ibid.) suggest that a product routing
analysis should then be undertaken and the selection based
upon the desired pathway through existing resource
centres.
In addition to this focal value stream selection issue, a
review of the VSM literature reveals an exclusive focus on
operational performance data. For example, Rother and
Shook (1998) in their original description of the learning
to see technique, and other authors who describe its general
application such as Bicheno and Holweg (2008), stress the
use of data such as process cycle time, changeover time,
uptime and production lead time. In their application of this
technique within an electronics manufacturing firm,
Worley and Doolan (2006) highlight inventory,
manufacturing lead time, quality, flexibility and customer
satisfaction as focal measures; whilst Abdulmalek and
Rajgopal (2007) in their VSM application within a steel
mill stress the measurement of production lead time along
with inventory at various stocking points within the entity.
In all of the VSM techniques identified during the review
exercise, there is a notable lack of a financial data. For
example, inventory data is typically captured as a number
of parts, and then sometimes converted into the ‘daysworth’
of consumption that this represents, but its monetary value
is not recorded. Likewise, there is, for example, no
evaluation of the additional revenue generated, costs
reduced nor return on investment as an outcome of the
project. There is consequently no financial dimension to
either the [above] focal value stream selection criteria, the
targeting of the associated improvement intervention
projects, or the evaluation of the subsequent impact of such
projects.
3. Methodology
3.1. Research objective
Given the characterisation of the literature in the previous
section, the following research objective was established to
guide this study:
To design, implement and assess a targeting method for
prioritizing and selecting a lean-type process improvement
intervention project that yields the largest financial impact
within a large and geographically dispersed manufacturing
operation.
3.2. Research strategy
The case study was adopted as the basis for addressing the
research objective detailed above because of our desire to
conduct an empirical enquiry that sought to explain a
contemporary phenomenon in a real-life context using
multiple sources of evidence (Yin 2003). The case study
can provide rich knowledge of a specific context (Meredith
1998; Yin 2003; Sousa and Voss 2008) and has a heritage
within both operations management and logistics where it
has been employed for research purposes that include
exploration, theory building, theory testing and theory
extension (Eisenhardt 1989; Ellram 1996; Voss,
Tsikriktsis, and Frohlich 2002).
AR is a variant of case research (Coughlan and Coghlan
2009; Brown and Vondracek 2013) and similarly offers the
advantage of yielding richness of insight over issues that
actually matter to practitioners (Eden and Huxham 1996).
The origins of AR are generally attributed to the work of
the psychologist Kurt Lewin in the 1940s (Eden and
Huxham 1996; Coughlan and Coghlan 2002; Hendry,
Huang, and Stevenson 2013); and particularly Lewin
(1946, 1947). However, since its inception, AR has
evolved to encompass a number of related approaches such
as action learning, action science, action enquiry and
participatory AR (see Eden and Huxham 1996; Checkland
and Holwell 1998). Given this evolution, within our paper
we adopt the interpretation of AR suggested by Platts
(1993, 9);
… the researcher not only participates in the activity but
seeks to direct and influence the way in which the activity
is conducted. He imposes his conceptual frameworks on
the tasks and interprets the events within these
frameworks. He is not so much concerned with gaining a
better understanding of current approaches to tasks as with
changing those approaches and observing the effects.
An AR approach was therefore adopted because the
researchers needed to actively participate with
practitioners from VehicleCo to influence all the stages of
the planned intervention project, observe its outcomes and
also to consider the general implications of this study
beyond this case firm (Gummesson 2000; Coughlan and
Coghlan 2009; Hendry, Huang, and Stevenson 2013).
Whilst this approach was deemed the most appropriate
to address the research reported upon within this paper, AR
is not without its criticisms. In concert with the case study
with which it shares common methodological
characteristics, these criticisms concern methodological
rigour and chiefly originate from a positivistic world view,
revolving around the validity of generalising to a wider
population from a [case] ‘sample size of one’ and the lack
of repeatability of such studies (Eden and Huxham 1996;
Checkland and Holwell 1998; Coughlan and Coghlan
2002). Premised upon such a world view, Gummesson
(2000), for example, asserts that case-based research can
generate hypotheses but not test them. This deems such
research insufficient for either generating or testing theory.
By contrast Eden and Huxham (1996, 80) recognise that
each AR intervention is highly situational, making AR an
unsuitable vehicle for theory testing. However, they argue
that AR will generate
… an emergent theory, in which the theory develops from
a synthesis of that which emerges from the data and that
which emerges from the use in practice of the body of
theory which informed the intervention and research
intent.
They likewise recognise that tools, techniques, methods or
models can form the basis of such theory. To support this
latter view and counter the concerns raised above, authors
have stressed the importance of deliberate research design
that considers the wider implications of the research and
effective roles and relationships within AR project teams
(Coughlan and Coghlan 2002; Hendry, Huang, and
Stevenson 2013). It is to these issues that we now turn.
3.3. Research design, data collection and analysis
Drawing upon Lewin’s (1946) work, numerous authors
have categorised the steps involved in the design of an
effective AR study, which is typically premised upon
iterative intervention/ reflection cycles. Perhaps, one of the
most influential of these techniques is the work of
Coughlan and Coghlan (2002), which was used as the
framework for the design of our study.
Due to its instrumental role within this study, the
following provides a brief summary of Coughlan and
Coghlan’s guidance on the purpose of each of these steps.
Table 1 details the main fieldwork-related activity that was
undertaken during each of these steps, and hence illustrates
how this guidance was translated into practice during this
first AR cycle at VehicleCo.
According to Coughlan and Coghlan (2002), the first
step of an AR cycle is a pre-step entitled Context and
Purpose. The aim of this is to clearly establish and
articulate the rationale for the proposed action by
addressing the question of why this project is necessary or
desirable from the perspective of the organisation.
Likewise, to also consider the rationale for undertaking the
research by addressing the questions: why is this project
worth studying, why is AR an appropriate methodology for
it, and what is its expected contribution to knowledge? For
the VehicleCo project, this step therefore entailed the
research objective developed earlier, along with the
concomitant practical terms of reference such as roles and
responsibilities, resource availability and time frames. A
steering group was then formed that included the heads of
VehicleCo’s engineering, manufacturing and accountancy
departments. Because of the nature of the research
objective, the project was ‘championed’ by the head of
finance. His role was to act as the main point of contact
between the steering group and project team that was
subsequently formed to deliver the project’s research
Table 1. Steps and activities in the VehicleCo AR cycle.
Step
Fieldwork activities
1. Context and
purpose
•
•
•
•
Establish the project’s research objective and practical terms of reference.
Form the project steering group.
Form the project team.
Initiate company-specific document collection.
2. Data gathering
•
•
•
•
•
Project team workshop to establish a block diagram of the high-level structural features and KPIs of
VehicleCo’s operation.
Follow-up interviews with attendees and shop floor staff to validate and elaborate upon the efficacy of
this block diagram.
Data mining of time series data from the ERP system to deepen insight into the firm’s operational and
financial performance.
Ongoing observation of shop floor practices.
Ongoing unstructured interviews with shop floor staff.
3. Data feedback
•
Miscellaneous ongoing feedback sessions to the project champion at the end of each pre-planned
monthly 2–3 day fieldwork event.
4. Data analysis
•
•
Analysis of numerous trends and relationships within the mined ERP time series data to establish what
the data inferred about the working principles, practices and subsequent performance within the firm.
Synthesis of four new key constructs: Profit & Loss Improvement Sensitivity Analysis, Inventory
Trend Analysis, Inventory-Working Capital Analysis and Inventory Provision Analysis.
5. Action planning
•
•
Construct a progress report and supporting presentation.
Formal presentation to the steering group.
•
Project team workshop to design the format of a new process improvement targeting mechanism,
subsequently entitled Big Picture Financial Map, and populate this with requisite data for VehicleCo.
Develop a business case and supporting presentation for the intervention project suggested by the
BPFM targeting mechanism.
6. Implementation
•
7. Evaluation
•
•
Formal presentation to the steering group.
Decision to execute the next AR cycle (design and implement the intervention project identified by the
BPFM).
objective. This project team contained three of the authors
of this article, who were acting as facilitators and were all
experienced action researchers. It also contained a group of
six mid-level managers drawn from and representing the
key fabrication and final assembly work centres found
within VehicleCo’s UK operation. Once the project team
was formed, they were tasked with collecting background
information such as company structure, hierarchy and
promotional literature, and hence initiating companyspecific (secondary) document collection.
The second step in the framework is entitled Data
Gathering. For the action researcher, data generation is
derived via active involvement in the day-to-day
organisational processes and interventions relating to the
AR project. This involves the collection of ‘hard data’ such
as operational statistics and financial accounts, and also
involves the collection of ‘soft data’ gathered via
observation, discussion and interviewing. Drawing upon
the previously collected background information, this step
was initiated at VehicleCo with the full project team
engaging in a highly interactive one-day workshop. The
purpose of this was to provide the newly formed team with
a common understanding and overview of the whole of
VehicleCo’s UK operation and the nature of the
management control information used within it, as none of
the team members currently had that requisite knowledge
due to the functional nature of their roles. The workshop
entailed the collective production of a simple block
diagram to highlight its main high-level structural features
and the main key performance indicators (KPIs) used
within the operation. The practitioners were tasked with
bringing along copies of relevant supporting evidence such
as work centre performance management documentation,
with the event itself involving free flowing discussion. In
the weeks following this workshop, the researchers
conducted a number of follow-up semi-structured
interviews with the attendees and salient shop floor staff.
The purpose of these interviews was to clarify and validate
a number of details raised during the workshop event, and
also to collect further information on the most influential
KPIs in use within the firm.
This information was subsequently added into the
block diagram drafted during the workshop, then circulated
to the team members for their information and further
confirmatory feedback. In order to deepen their insight into
the firm’s current working principles and practices in
addition to quantifying the firm’s resulting operational and
financial performance, the project team next undertook a
large-scale data mining exercise of data for the previous 7
months. This was conducted with the active participation
of the VehicleCo accountancy department and involved
obtaining data such as the physical location and value of
all work in progress (WIP) per contract, monthly operating
expenses and sales revenue per contract. This time series
was extracted from the firm’s enterprise resource planning
(ERP) system into a series of Excel spreadsheets for
subsequent analysis.
Data Feedback, the third step in the AR cycle, involves
the researcher feeding back the gathered data to the client
organisation with the view to making this data available for
analysis. Research is rarely conducted in a neat linear–
sequential manner, and this project was no exception.
During the VehicleCo study the data gathering and
feedback steps were conducted as a series of iterative minicycles. This took the form of pre-planned monthly 2–3 day
fieldwork events conducted over a 6month period, with
each such event culminating in an informal feedback
session to the project champion. An agenda was agreed in
advance of each event, and each involved the full-time
participation of all nominated project team members. The
fieldwork event itself typically comprised a number of
formally minuted meetings as well as structured and
unstructured fieldwork activity. During the course of both
this step and the analysis step that followed, ongoing
observation of shop floor practices and unstructured
interviews with shop floor staff was afforded by the
richness of access and location of the project team’s office
space at a central location within the operation. The
observation was supplemented by digital photography to
evidence salient practices. Each fieldwork event
culminated in minuted supplementary data collection tasks
to be completed by the practitioners in advance of the next
scheduled whole-team event. As a consequence of these
procedures, the evidence base of our project was
significant. It amounted to over 50 h of interviews, 20 h of
observation and hundreds of pages of company-specific
documents, digital photographs and ERP data extracts.
Data Analysis involves the identification of issues and
subsequent focus for action. Importantly, Coughlan and
Coghlan (2002) assert that the criteria and tools for analysis
should be directly linked to the purpose of the research and
the aim of the interventions. Of critical importance is the
requirement for such data analysis to be a collaborative
exercise, involving both the researchers and members of
the client organisation. During the VehicleCo study this
step was a significant undertaking, being conducted over a
further 6-month period. As per the above, this took the
form of monthly 2–3 day collaborative fieldwork events
undertaken at VehicleCo’s premises, with each such event
culminating in feedback to the project champion. The
analysis stage of an AR project typically entails the coding
and content analysis of interview data (see Lavikka,
Smeds, and Jaatinen 2009; Brown and Vondracek 2013).
However, because of the nature of the research objective,
the analytical focus at VehicleCo was on the hard data
extracted from the firm’s ERP system to establish what it
could further infer about the working principles and
practices deployed by the firm, and the resulting
performance levels attained. Therefore, during this period,
this time series data was first smoothed using a standard
30-day month in order to facilitate inter-month
comparisons. It was then manipulated at the factory, work
centre, department, resource (a machine or group of like
machines) and item level. The exercise was approached
with an open mind and no preconceived ideas regarding the
specific relationships to explore, although particular
attention was paid to inventory relationships as these
would provide insight into how the firm was deploying its
capacity. This resulted in the identification of numerous
trends and relationships that had not previously been
explored within the firm such as sales and WIP, sales and
working inventory (WIP + finished goods) and factory
hours worked and WIP. The product of this exercise was
the synthesis of four new constructs that were believed to
be salient for the targeting of the consequent improvement
intervention; a Profit & Loss (P&L) Improvement
Sensitivity Analysis, Inventory Trend Analysis, InventoryWorking Capital Analysis and Inventory Provision
Analysis.
The Action Planning step entails translating the insight
gained during the previous step into a practical intervention
plan, and it is again important that this be a collaborative
exercise. At VehicleCo, this step was initiated by the
project team preparing a progress report and supporting
presentation that encapsulated the above insight and
suggested project next steps. A team presentation was
made to the steering group, who subsequently gave their
approval for the project to progress.
Implementation involves the client organisation
enacting the above planned actions. A second wholeproject team workshop was consequently convened at
VehicleCo. The purpose of this was to collaboratively
design the new process improvement targeting mechanism
itself. Taking the block diagram developed during the data
gathering stage as a starting point, the format of this new
mechanism was duly innovated and subsequently labelled
a Big Picture Financial Map (BPFM) by the team. Drawing
upon the previously collected data, and again with the
active involvement of the accountancy department, this
BPFM was then populated with the salient data from a
representative month within the firm’s commercial
operation. Its subsequent analysis suggested the logical
point of process improvement intervention within
VehicleCo’s dispersed UK operation. Based upon this
insight, this step culminated in the development of the
business case and supporting presentation for this
suggested intervention project.
The seventh, Evaluation step, entails assessing both the
intended and unintended intervention outcomes with
regard to the stated purpose for the AR project. The project
steering group is the ultimate arbiter of this evaluation
exercise. During the VehicleCo project, this step started
with the formal presentation of the above business case to
the steering group in order to seek approval to plan and
execute the intervention project. After due consideration
and a significant question and answer session with the
project team, the steering group assessed that this new
BPFM mechanism and the intervention project that it
suggested indeed had the potential to yield a significant
financial impact within the firm. They consequently
approved this intervention, hence initiating the second
intervention/reflection AR cycle at the firm. The detailed
characterisation of this 12-month project is outside the
research objective reported upon within this paper.
However, during the Evaluation step of this second AR
cycle, the financial impact of the targeted intervention was
independently valued by the VehicleCo accountancy
department to be worth at least £850K per annum to the
firm. This impact therefore validated the steering group’s
assessment decision.
The final step in Coughlan and Coghlan’s (2002) AR
cycle is entitled Monitor. This is categorised as a ‘metastep’
by the authors, as monitoring should ideally occur during
each of the previous steps. This step is the preserve of the
researchers, who in addition to a concern for evaluating the
practical outcomes (above), also have a concern for
reflecting upon the efficacy of the process by which these
outcomes are achieved. Because of the inherent nature of
this step, it is omitted from Table 1. However, to support
reflective practice, a journal containing ideas, thoughts and
observations was maintained by the researchers. This was
supplemented by a comprehensive project file that
contained copies of all company documentation, agendas,
minutes and correspondence collected during the project.
The insight provided by this exercise was utilised to good
effect in the subsequent AR cycle.
3.4. Data validity
As indicated earlier, AR is a variant of case study research
and the validity of the data collected during AR can be
evaluated against the criteria applied to case research
(Brown and Vondracek 2013). Because of the nature of
such research, the issues of validity and reliability are
highly related. Yin (2003) suggests four tests of validity in
addition to the tactics required to realise them, and these
offer a yardstick by which our VehicleCo study can be
evaluated: construct validity, internal validity, reliability
and external validity.
Construct validity necessitates establishing the correct
operational measures for the phenomenon being studied. In
accord with the tactics suggested by Yin, we used multiple
sources to establish ‘chains of evidence’. All the collected
data was reviewed for accuracy by the source informants
and practitioners on the project team, with the steering
group presentations forming a further plausibility test.
Importantly, we also employed methodological
triangulation to generate complementary data (after Denzin
1970). Drawing upon the discussion in the previous subsection, Table 2 summarises the instruments that we used
per step in the VehicleCo AR cycle, with the workshops
and ERP data mining exercise being, respectively,
deconstructed into the categories ‘interviews’ and
‘company-specific document collection’.
The internal validity of our study was underpinned by
the degree of collaboration and research access that was
afforded the researchers, along with the nature and extent
of dialogue with the project champion. This was enhanced
via the application of time series data analysis. Yin’s third
test (reliability) refers to an ability to demonstrate that the
operations of a study can be repeated with the same results,
and it is to address this criterion we have detailed the
research design, data collection and analysis procedures
within this article. Yin’s fourth and final test of external
validity refers to the domain in which the study’s findings
can be generalised. Our study suffered the inherent
limitation of being a single, AR case and therefore details
one particular application. We will return to this limitation
in the Conclusion of our paper, where we outline planned
future directions of research for increasing the external
validity of this study.
3.5. Confidentiality
Under the terms of a strict confidentiality agreement, a 4year moratorium on publication has been observed for this
study. Other measures have also been applied within this
paper to assure the anonymity of the firm whilst
simultaneously maintaining the integrity of the findings.
These measures include the use of the alias VehicleCo for
the case firm, and the disguise of all terminology that could
be used to identify it. This includes all specifics regarding
the firm’s product portfolio, the industry sector within
which it operates, its geographic location and all reference
to its annual turnover and scale of employment. Lastly, all
financial and operational data has been disguised by means
of a constant modifying factor.
parts of the business, and also identified the main plants,
work centres and inventory holding points within each of
these. Likewise, the main production lines (assets that are
dedicated to specific customer contracts) were identified
within the final assembly area. Whilst the format of the
diagram was unremarkable, the process used by the project
team to produce it rapidly established that comprehensive
cost data was readily available within VehicleCo, and at
factory, work centre, department, and resource level.
Operational performance measurements were found to also
reflect this trend. However, information on lead times
through various parts of the operation was difficult to
establish, although cycle time and quality/scrap
information at individual resource level was again readily
available.
During the course of the initial project team workshop
and the interviews that followed it, a concerted effort was
made to establish the main KPIs that were used for
management control purposes within the operation. A large
number of different metrics were duly identified, and the
most influential of these were annotated onto the block
diagram to act as an aide-memoire. These metrics were
found to be almost universally of a unit cost type and
included controllable overhead rate, controllable wrap rate
and direct labour rate. Measurements influence behaviour
Table 2. Data collection instruments used during the VehicleCo AR cycle.
Data collection instruments
AR cycle step
1. Context and
purpose
Interviewsa
Observationb
Company-specific document
collection
Reflective journal and project
file
✓
✓
2.
3.
Data gathering
Data feedback
✓
✓
✓
✓
✓
4.
5.
Data analysis
Action planning
✓
✓
✓
✓
✓
6.
Implementation
✓
7.
Evaluation
a
Includes unstructured and semi-structured interviews.
Includes digital photographs.
b
4. Discussion
As indicated in the previous section, the initial steps of the
AR cycle at VehicleCo involved forming the project team
and establishing among them a common, high-level
understanding of the structure of the firm’s UK operation.
The main deliverable was the high-level block diagram that
distinguished between the final assembly and fabrication
✓
✓
(Hrebiniak and Joyce 1984; Neely, Gregory, and Platts
1995) therefore unsurprisingly the follow-up interview
evidence and subsequent secondary analysis of the
operational performance data established that for many
managers such unit cost measures were being interpreted
as ‘make as much as you can in the time available’, a
culture of local optimisation.
Following the data feedback step, the team turned to the
data analysis of the financial and operational data contained
in the firm’s ERP system. Data was extracted, smoothed
and then combined in ‘interesting’ ways to establish
whether any trends or relationships could be discerned that
would cast insight into the working principles and practices
that were in common currency within the firm. Whereas a
number of such relationships were explored, only two are
presented here due to space constraints, but suffice to
illustrate the general findings.
The first of these is the relationship between sales
revenue and WIP (Figure 1). The data on both of the axes
has been smoothed, and plotted over a representative 7month period. WIP equates to manufacturing lead time, so
the floor as orders decline. This suggests that perhaps the
plant is being kept busy for the sake of being busy, in an
attempt to maintain utilisation figures and unit costs.
The second exploratory relationships are illustrated in
Figure 2. This chart reports upon the same 7-month period
and illustrates the relationship between total factory hours
worked and the total WIP value. Both April and July are
notable holiday periods within the locale of the plant, and
it is interesting that the chart seems to indicate extra hours
worked and the building of WIP in the months adjacent to
Figure 1. Relationship between sales and work in progress.
Figure
2.
Relationship
between factory hours worked and work in progress.
the following chart illustrates a relatively long average lead
time of 80 days plus, although this is not unusual for the
industry concerned. The most interesting feature of the
diagram is that during the period May– July, WIP and
hence lead time grows in an inverse relationship to the sales
velocity. It seems that more material is being released to
each of these two periods. This reinforces the pattern that
emerged in the previous figure. These two examples, taken
in conjunction with the wider accompanying trends and
relationships explored during this part of the study, clearly
indicated that the VehicleCo operation was working to
conventional mass production principles and practices
whose lineage could be traced to scientific management
(Taylor 1911), such as the importance of local rationality
and scale-derived efficiency.
Having gained this insight, the project team now turned
to establishing a greater depth of understanding of the
effects of these working practices on the current financial
and operational performance of the firm. Whilst more
extensive analysis was performed as indicated earlier, four
new constructs provided particular insight. These were a
significant impact on the bottom line. The cost categories
are arranged in a descending order of impact of this 5%
improvement.
Figure 3 therefore illustrates that the initiative that
would yield the largest financial impact to the firm would
be an across-the-board increase in the sales price of its
product offerings. This was deemed impractical by senior
management given the state of the market at that point in
time, which was highly competitive and in a period of slow
P&L Improvement Sensitivity Analysis, Inventory Trend
Analysis, Inventory-Working Capital Analysis and
Inventory Provisions Analysis.
Figure 3 illustrates the P&L Improvement Sensitivity
Analysis, which represents an important innovation for the
targeting and development of a business case for a leantype process improvement intervention. The lefthand side
of the figure illustrates a conventional P&L statement for
the previous 12-month period up to that census month to
date. The data used for the production of this statement
were extracted directly from the firm’s accounting system
and utilised its conventional cost category labels. The only
unconventional category is ‘throughput’, which is a
throughput accounting (TA) construct defined as sell price
– material cost (Goldratt and Cox 1984; Corbett 1997).
This P&L statement therefore summarises the real money
that the company generated in terms of sales in that period.
It also summarises
recovery from recession. In any case, VehicleCo received
a transfer price from the parent enterprise and did not set
the end price of its product offerings to the final customer.
The second largest financial impact would be delivered via
an across-the-board reduction in the cost of materials.
Again, this would entail negotiation with external
stakeholders, and was deemed impractical due to the nature
and number of contracts concerned. However, the third
largest opportunity lay entirely within the internal control
of the firm itself. This was the potential offered via a
programme for increasing throughput, because a 5%
improvement would add £14.28 million to the bottom line.
This is nearly 273% larger an impact than the ‘next best’
opportunity, the targeting of indirect salaries. In fact, the
throughput improvement opportunity represented 162% of
the bottom line impact of the combined reduction of both
indirect and direct salaries.
Having obtained this preliminary financial insight, the
second task was to analyse the firm’s total inventory
holdings over the last 6 months of the period reported upon
above (Figure 4). Further data was extracted from the
firm’s ERP system for each of the raw material (RM), WIP
and finished goods (FG) categories of inventory. As per the
daysworth figures presented in the earlier diagrams, this
data was then smoothed using a standard 30-day month to
establish an average ‘daysworth of sales coverage’ figure
Figure 3. Profit & loss improvement sensitivity analysis (5%).
the real money it spent to realise that revenue. The
righthand side of the figure forms the sensitivity analysis.
It illustrates the financial impact of a 5% improvement in
each of the respective cost categories. The reason for
selecting 5% was that this figure was considered a modest
improvement goal by the firm, yet it still revealed a
for each inventory category for each of the 6 months
concerned. By the way of illustration, the equation used to
establish the ‘RM daysworth of sales coverage’ figure for
February was: (raw material value at end of February/raw
material cost of sale for March) × 30 days. Figure 4
illustrates relatively stable inventory coverage for each of
the three inventory categories over the 6-month period
reported upon. The following average sales coverage
figures were also established: RM (21 daysworth), WIP (83
daysworth) and FG (5 daysworth). It is notable that the FG
figure is a function of spare parts rather than whole
bodywork or auxiliary structures, as the value and nature
of these products ensured their rapid distribution to the
downstream customer. The levels of WIP witnessed
seemed to be a structural feature of the industry within
which the firm operates, which is typified by penalty
clauses for late delivery. However, the working practices
established earlier were also a likely contributory factor.
After establishing the above average inventory
coverage figures, it was possible to utilise them to gain
further insight into the working capital implications of the
current state of the VehicleCo operation (Figure 5).
Inventory equates to manufacturing lead time (MLT), as
indicated earlier. Therefore, when the average RM, WIP
Figure 4. Inventory trend analysis.
Figure 5. Inventory-working capital analysis.
and FG inventory coverage figures are added together, they
yielded a derived MLT of 109 days. A check was made
with production staff to validate the reliability of this figure
as an approximate to the actual average MLT witnessed in
reality. Based upon a standard 30-day term for debtors to
pay for their FG, this then equates to a 139-day
manufacturing-to-cash lead time, from receipt of the RM
by VehicleCo in order to start the production process
through to receipt of payment from the customer for the
resulting FG. Payable terms and conditions for
Figure 6. Inventory provisions analysis (£000s).
goods received by VehicleCo were 50 days. It was
established that September was a representative month and
that the September month-end inventory valuation figure
was £168 million. The firm was therefore being called upon
to finance this value of inventory for 89 days. This is
clearly a significant financial burden.
The fourth and last of the new constructs produced
during this stage of the study was an inventory provisions
analysis (Figure 6). Initial comments made by a number of
staff during the confirmatory feedback interviews in the
data gathering step had suggested that write-off costs were
not significant within the firm. This new construct was
therefore made in order to substantiate these claims, and
the results were surprising. The accountancy department
again assisted with the provision of the necessary data. It
was subsequently found that there were eight inventory
provision categories in operation, and that £3.4 million had
been written off in the first 9 months of the year to date.
This data, taken in conjunction with the follow-up
interviews with the accountants and managers in the
operational areas concerned, reinforced the fact that
inventory was relatively slow moving and that engineering
changes were relatively frequent due to the nature of the
product portfolio. These figures therefore suggested further
evidence that resource centres throughout the firm were
optimising the organisation of their work to respond to the
unit cost nature of their prevailing KPIs, scheduling batch
sizes to maximise machine utilisation rather than flow
rates, and hence exposing the surplus stock to the risk of
obsolescence due to engineering changes.
In summary then, the analysis of the financial and
operational data yielded by the four key new constructs
developed during this step of the AR cycle clearly indicates
that the most logical course of action for this firm, given its
current level of performance, would be to increase
throughput. To achieve this, it would be necessary for
VehicleCo to reduce its MLT in order to improve the flow
of work through the production process. Reference to
Figures 5 and 6 reveals that WIP represents by far the
largest component of the MLT within this case operation.
Unlike RM and FG, which are generally a reflection of the
variability of supply and demand (respectively), the
amount of WIP a firm holds is a function of its policies and
the way that it deploys its capacity. The amount of WIP is
therefore entirely within the control of the firm. The data
therefore suggest that the focus for the intervention should
be the systematic reduction of MLT via the reduction of
WIP. This course of action would yield a number of
benefits. Of immediate impact would be the reduction in
production lead time, which would be a function of the
amount of WIP (daysworth) reduced. It would similarly
reduce working capital requirements and concomitant
inventory management costs. The increased inventory
turns that would result from this exercise would likewise
reduce the exposure to risk, and hence write-off costs, that
are associated with engineering changes. Lastly, improved
throughput could also reasonably be expected to translate
into improved due date performance. A WIP reduction
initiative of this sort therefore promises more than
unfocused waste reduction of the type characteristic of
many lean initiatives (Bicheno and Holweg 2008). It
instead offers both genuine cost reductions to the producer
and improved utility to the downstream customer.
After presenting the previous findings to the steering
group, the project entered the Implementation step of the
study. It was during this step that the new mechanism was
to be designed to help identify and target the most
appropriate location for this WIP reduction intervention
within VehicleCo’s large and geographically dispersed
manufacturing operation. ‘Appropriate’ in this context was
interpreted to mean the potential for the largest financial
impact within the firm. Consequently, to inform this
exercise additional analysis was undertaken again in
concert with the VehicleCo accountancy department, and a
second whole-project team workshop was convened. First,
the format of this new mechanism was developed and
agreed upon. It was then populated with data for a
representative (commercially typical) month. This new
mechanism came to be named a BPFM and it took the guise
of a new form of a high-level schematic representation of
the financial current state of the whole case operation at a
representative point in time (Figure 7).
An important feature of this BPFM is that it embodies
the three TA measures of throughput, inventory and
operating expense, proposed by Goldratt and Cox (1984)
as part of their TOC. According to these authors, the goal
of any private sector organisation is to make money, and
they introduce the concept of TA to assist managers in
making decisions that attain this. TA is premised upon
three metrics that differ in definition to the more
conventional usage of the terms (Corbett 1997):
‘Throughput’ is the rate at which the system generates
money through sales. This is the volume of sales expressed
in terms of money rather than units, and products only
become throughput when they are actually sold. The
second measure is ‘inventory’, which is defined as the
money invested by the system in things that it intends to
sell. Within TA, this measure encompasses equipment and
facilities in addition to the conventional inventory
categories of raw material, WIP and finished goods.
However, in the case of the latter categories, it equates to
the direct material cost and excludes any notion of value
added during the production process. The final TA measure
is ‘Operating Expense’, (OE) which is defined as the money
that the system spends to convert inventory into
throughput. The TA conception of operating expense
adopts a marginal costing approach and notably excludes
all overhead allocations (Boyd and Gupta 2004). Goldratt
and Cox (1984) argue that all management decisions
should be based upon these three criteria alone, with the
aim being to increase throughput whilst simultaneously
decreasing inventory and operating expense. Such a
decision is classified as ‘productive’ as it will contribute to
the organisation achieving the goal.
The BPFM distinguishes between the raw material,
fabrication and final assembly areas of the VehicleCo
manufacturing operation. Within the final assembly area,
all of the main product line contracts (value streams) are
identified along with the total throughput associated with
each in the period concerned. This captures the real money
flowing into the firm’s bank account. Within the fabrication
area, all of the firm’s main machining work centres are
identified, along with the major material work flows
between these centres and the final assembly areas. Each
such centre is represented as a box, with its monthly OE
displayed at the bottom. This is the sum of the fixed costs
that are incurred in running the centre for that period,
regardless of its production volume, and it excludes any
overhead allocations as indicated above. The BPFM also
captures the WIP inventory material and utility cost
valuation that is tied up at each major stock point in the
system. However, absolute inventory value doesn’t tell us
much about whether this is too much, or too little, for the
demands placed upon it. Each such value is consequently
converted into a daysworth of sales coverage figure to cast
further insight into this issue, using the calculation detailed
earlier.
The analytical work conducted to produce the BPFM
yielded three notable findings. The first of these was the
triangulation of the average MLT figure established during
the earlier data gathering step. A second finding was that
nearly half (47%) of the firm’s total labour was to be found
in the upstream fabrication work centres as compared to
53% in the final assembly area. This was important because
it meant that the firm’s established internal lean initiative,
which was focusing on labour cost reduction in final
assembly, was in fact currently targeting only
approximately 50% of its potential scope. When
considered in conjunction with the insight gained via the
P&L Improvement Sensitivity Analysis (Figure 4), this
also meant that a company-wide improvement in
throughput offered an impact potential of nearly 600%
more than this established programme of labour cost
reduction, and not 300% as originally estimated.
The third finding was the identification of the most
appropriate location to intervene within the complex case
operation. Figure 7 reveals that the fabrication work centre
that has the highest WIP in terms of absolute inventory
value is WC5-Hybrid Fabrications. It also has the highest
OE. However, the BPFM likewise revealed that the centre
that was carrying the highest inventory in terms of
daysworth of sales coverage was WC2, where the
individual bodywork and auxiliary panels were produced.
In addition, this centre represented the third highest
operating expense figure. Unlike WC5 which was
perceived as an exemplary centre and was actually
Figure 7. Big picture financial map of VehicleCo’s UK operation.
producing ‘early’ approximately 60% of the time, WC2
was considered an operational ‘problem child’. It was
suffering from high rework and scrap rates and long lead
times. It was also suffering from poor delivery schedule
adherence, being late approximately 60% of the time.
This was evaluated to be causing on average six jig
stoppages per week within the final assembly area.
Hence, WC2 was already causing much disruption to the
production schedules of its downstream final assembly
customer.
This comparative performance requires brief
explanation. Whilst Figure 7 illustrates that WC2 feeds
WC5, this workflow represents a relatively small number
of parts. The products of these two work centres are
produced in response to independent demands from the
final assembly lines. WC5’s early performance profile
merely represents another form of poor delivery
schedule adherence. For example, final assembly
demands 50 of WC5’s component X products to be
assembled into its final composite structures. However,
100 component Xs are produced as a consequence of the
local batching rules and efficiency metrics within WC5.
This is recorded on the company’s performance
measurement system as ‘50 × early’. This is consequently
a sequencing, batching and performance measurement
issue.
Returning to WC2’s issue and during the period on the
run up to the presentation of the study’s results to the
steering group, the firm’s order book took a significant
uplift as the market started to recover. When the firm
modelled the implications of this new order book, they
found that within the immediate months ahead, they would
be encountering a serious operational problem within
WC2. In accord with conventional mass production
practice, the traditional response to more demand had been
to push more work into the system; therefore, tying up
more working capital, producing more WIP and
consequently more risk of obsolescence and damages (but
not more throughput). The problem was poised to become
acute over the coming period as, apart from the cost and
working capital implications, the uplift in the order book
meant that if current working practices were followed,
there was not going to be enough physical space within
WC2 to accommodate all of the panels that were forecast
to be produced and moved around this centre.
Consequently, WC2-Panels was identified as the most
appropriate location for the planned process improvement
intervention.
These recommendations were accepted by the steering
group, stimulating the second AR cycle within the firm.
The resultant 12-month process improvement project
within WC2 is detailed in Darlington et al. (2015). This
intervention took the form of a Drum Buffer Rope (DBR)
type of pull system (after Goldratt and Cox 1984) and was
implemented with significant operational and financial
success. It resulted in a 23-day (56%) MLT/ WIP reduction
that was independently valued by the VehicleCo
accountancy department to be worth at least £850K per
annum. The improved throughput translated into an
increase in WC2’s inventory turns from 9.1 to 21.2 per
annum, and the improved lead time and reliability resulted
in jig stoppages in final assembly falling to less than one
per week. As a consequence of this economic impact, the
WC2 DBR project was nominated by VehicleCo for the
parent enterprise’s annual worldwide process improvement
competition that year, where it won the first prize.
5. Conclusions
The industrial problem that formed the context for this AR
study was the need for an objective method for targeting
(prioritising and selecting) lean intervention projects
within the case firm; one that encompassed financial as well
as operational criteria. However, a structured literature
review on this topic was unable to identify any decision
support mechanism for doing so. Instead, it established that
the kaizen technique that forms the medium for process
improvement within the lean paradigm has evolved over
the last three decades. In its original conception, kaizen
entailed continuous incremental improvement in response
to localised suggestions or workplace problems by
standing quality or KC teams. This has become known as
process kaizen. By contrast, the more contemporary flow
kaizen type entails the proactive design and
implementation of discrete lean process improvement
projects. It is this latter type of kaizen that is practised at
VehicleCo and which necessitates a supporting targeting
device. Flow kaizen seems to be becoming increasingly
prevalent, and this in turn appears to be driven by the
influence of various popular VSM techniques that have
emerged since the latter half of the 1990s. Indeed, VSM
has emerged as the de facto vehicle for implementing flow
kaizen-type projects. However, mirroring the gap
identified above, existing VSM techniques lack guidance
for conceiving or objectively selecting the value stream
that is to form the focus for the mapping exercise. They
likewise reveal an exclusive focus on operational
performance criteria, and contain no direct financial
dimension. This clearly has implications for both the risks
and returns associated with the resultant improvement
projects.
The development of the resulting targeting method at
VehicleCo involved the main body of evidence being
derived from extensive analysis of financial and
operational data extracted from the firm’s ERP system,
along with two structured workshops that involved
multiple informants drawn from the firm’s production work
centres and accountancy department. During this exercise,
a number of previously unexplored trends and relationships
in the data were investigated. These were used to establish
that the firm was operating according to conventional mass
production principles, such as the importance of local
rationality and scale-derived efficiency. Four key
constructs were then developed to deepen the team’s
understanding of the effects of these practices on the
financial and operational performance of the firm. This
analysis suggested that the most logical course of action for
the firm was to increase throughput via a targeted WIP
reduction initiative. A new targeting mechanism named a
BPFM was subsequently developed that drew upon the
information derived during the preceding stages of study,
and this was used to identify which of the work centres
within VehicleCo’s six UK plants offered the greatest
financial potential for the initiative identified. The resultant
intervention yielded a very significant financial and
operational impact at the case firm, and as a consequence,
it was awarded the first prize at the subsequent annual
worldwide process improvement project competition that
was held by the parent enterprise. This impact validated the
efficacy of the new targeting method and hence heralded
achievement of the research objective stated above.
As a consequence, it is possible to conclude that this
paper makes notable contributions to both the academic
and the practitioner. The academic contribution of this
paper resides in its addressing of the gap highlighted
above, by offering a financially driven method for
objectively prioritising and selecting a process
improvement intervention that was proven effective within
the case manufacturing environment. It is important to
stress that this new targeting method is not presented as a
substitute for the existing VSM techniques discussed in
this paper. It is instead presented as a complementary
method to be conducted as precursor to the deployment of
such techniques. This new method itself features a number
of innovative constituent techniques that include four new
constructs for deepening the understanding of the working
principles and financial performance of the focal case firm,
and also a new technique for mapping the financial current
state of the focal firm.
The practical contribution of this study for the future
economic impact of the lean programme at VehicleCo is
self-evident. If this targeting method proves to be
generalisable, it could likewise offer improved economic
impact potential for the firms adopting it to support their
indigenous lean process improvement programmes.
However, the limitations of the new method’s external
validity need to be recognised as it was derived from a
single AR case. As VehicleCo was purposively selected, it
is possible to conclude with a degree of confidence that the
external validity of our results will extend to similar
contexts (Silverman 2000). Therefore, the new targeting
method is likely to prove effective in other large, multi-site
manufacturing operations. Conversely, the more different
a new context to that of our source AR case, the more
unreliable will be the transferability of our results.
Adopting Yin’s (2003) advice to apply replication logic to
increase external validity, future research will be
conducted in a similar, large manufacturing operation. To
improve empirical generalisation, future research will also
be conducted in a variety of new contexts (Lavikka, Smeds,
and Jaatinen 2009). Foremost amongst these contextual
selection criteria will be scale, manufacturing process
technology and operational complexity. In each such
replication study, Platt’s (1993) three criteria for
assessment will be considered: feasibility (could the new
method be followed); usability (how easily could the
method be followed); and utility (did the method derive the
expected outcomes)?
Acknowledgement
The authors are grateful for the research access and time provided
by the staff at VehicleCo.
References
Abdulmalek, F. A., and J. Rajgopal. 2007. “Analyzing the
Benefits of Lean Manufacturing and Value Stream Mapping
via Simulation: A Process Sector Case Study.” International
Journal of Production Economics 107 (1): 223–236.
Bayou, M. E., and A. D. De Korvin. 2008. “Measuring the
Leanness of Manufacturing Systems: A Case Study of Ford
Motor Company and General Motors.” Journal of Engineering
& Technology Management 25 (4): 287–304.
Benders, J., and K. van Veen. 2001. “What’s in a Fashion?
Interpretive Viability and Management Fashions.”
Organisation 8 (1): 33–53.
Bicheno, J., and M. Holweg. 2008. The Lean Toolbox: The
Essential Guide to Lean Transformation. 4th ed. Buckingham:
PICSIE Books.
Boyd, L., and M. Gupta. 2004. “Constraints Management: What
is the Theory?” International Journal of Operations and
Production Management 24 (4): 350–371.
Brown, S., and P. Vondracek. 2013. “Implementing Time-based
Manufacturing Practices in Pharmaceutical Preparation
Manufacturers.” Production Planning & Control 24 (1): 28–46.
Checkland, P., and S. Holwell. 1998. “Action Research: Its
Nature and Validity.” Systemic Practice and Action Research
11 (1): 9–21.
Corbett, T. 1997. Throughput Accounting. Great Barrington,
MA: North River Press.
Coughlan, P., and D. Coghlan. 2002. “Action Research for
Operations Management.” International Journal of Operations
& Production Management 22 (2): 220–240.
Coughlan, P., and D. Coghlan. 2009. “Action Research.” In
Researching Operations Management, edited by C. Karlsson,
236–264. Abingdon: Routledge.
Dahlgaard, J. J., and S. M. Dahlgaard-Park. 2006. “Lean
Production, Six Sigma Quality, TQM and Company Culture.”
TQM Magazine 18 (3): 263–281.
Darlington, J., M. Francis, P. Found, and A. Thomas. 2015.
“Design and Implementation of a Drum-buffer-rope pullsystem.”
Production Planning & Control 26 (6): 489–504. Deming, W. E.
1986. Out of Crisis. Cambridge: Cambridge
University Press.
Dennis, P. 2007. Lean Production Simplified: A Plain Language
Guide to the World’s Most Powerful Production System. 2nd
ed. New York: Productivity Press.
Denzin, N. 1970. The Research Act: A Theoretical Introduction
to Sociological Methods. Chicago, IL: Aldine.
Duggan, K. J. 2002. Creating Mixed Model Value Streams:
Practical Lean Techniques for Building to Demand. New York:
Productivity Press.
Eden, C., and C. Huxham. 1996. “Action Research for
Management Research.” British Journal of Management 7 (1):
78–86.
Eisenhardt, K. M. 1989. “Building Theories from Case Study
Research.” Academy of Management Review 14 (4): 532–550.
Ellram, L. M. 1996. “The Use of the Case Study Method in
Logistics Research.” Journal of Business Logistics 17 (2): 93–
138. Feld, W. M. 2000. Lean Manufacturing: Tools, Techniques,
and How to Use Them. Boca Raton, FL: CRC Press.
Goldratt, E. M., and J. Cox. 1984. The Goal. Aldershot: Gower.
Groover, M. P. 2007. Automation, Production Systems and
Computer-Integrated Manufacturing. 3rd ed. Upper Saddle
River, NJ: Prentice Hall Press.
Gummesson, E. 2000. Qualitative Methods in Management
Research. 2nd ed. London: Sage.
Gurumurthy, A., and R. Kodali. 2011. “Design of Lean
Manufacturing Systems using Value Stream Mapping with
Simulation: A Case Study.” Journal of Manufacturing
Technology Management 22 (4): 444–473.
Hayes, R. H. 1981. “Why Japanese Factories Work.” Harvard
Business Review 59 (4): 56–66.
Hendry, L., Y. Huang, and M. Stevenson. 2013. “Workload
Control:
Successful
Implementation
Taking
a
Contingencybased View of Production Planning and Control.”
International Journal of Operations & Production Management
33 (1): 69–103.
Hines, P., M. Holweg, and N. Rich. 2004. “From Strategic
Toolkit to Strategic Value Creation: A Review of the Evolution
of Contemporary Lean Thinking.” International Journal of
Operations and Production Management 24 (10): 994–1011.
Hines, P., and N. Rich. 1997. “The Seven Value Stream Mapping
Tools.” International Journal of Operations and Production
Management 17 (1): 47–64.
Hrebiniak, L. G., and W. F. Joyce. 1984. Implementing Strategy.
New York: Macmillan.
Imai, M. 1986. Kaizen. New York: McGraw Hill.
Imai, M. 1997. Gemba Kaizen. New York: McGraw Hill.
Ishikawa, K. 1985. What is Total Quality Control? The Japanese
Way. Englewood Cliffs, NJ: Prentice-Hall.
Jones, D. T., and J. Womack. 2002. Seeing the Whole: Mapping
the Extended Value Stream. Brookline, MA: The Lean
Enterprise Institute.
Karlsson, C., and P. Ahlstrom. 1996. “Assessing Changes
Towards Lean Production.” California Management Review
39 (4): 118–151.
Krafcik, J. 1988. “The Triumph of the Lean Production System.”
Sloan Management Review 301: 41–52.
Lasa, I. S., R. De Castro, and C. O. Laburu. 2009. “Extent of the
Use of Lean Concepts Proposed for a Value Stream Mapping
Application.” Production Planning & Control 20 (1): 82–98.
Lavikka, R., R. Smeds, and M. Jaatinen. 2009. “Coordinating the
Service Process of Two Business Units Towards a Joint
Customer.” Production Planning and Control 20 (2): 135–146.
Lawler III, E. E., and S. A. Mohrman. 1985. “Quality Circles after
the Fad.” Harvard Business Review 63 (1): 64–71. Lewin, K.
1946. “Action Research and Minority Problems.” Journal of
Social Issues 2: 34–46.
Lewin, K. 1947. “Frontiers in Group Dynamics: Channel of
Group Life: Social Planning and Action Research.” Human
Relations 1: 143–153.
Lewis, M. A. 2000. “Lean Production and Sustainable
Competitive Advantage.” International Journal of Operations
and Production Management 20 (8): 959–978.
Liker, J. K. 2004. Becoming Lean: Inside Stories of US
Manufacturers. New York: Productivity Press.
Lyonnet, B., and R. Toscano. 2014. “Towards an Adapted Lean
System – A Push-pull Manufacturing Strategy.” Production
Planning & Control 25 (4): 346–354.
Martinez-Jurado, P. J., and J. Moyano-Fuentes. 2014. “Key
Determinants of Lean Production: Evidence from the
Aerospace Sector.” Production Planning & Control 25 (4):
332–345.
Melton, T. 2005. “The Benefits of Lean Manufacturing: What
Lean Thinking has to Offer the Process Industries.” Chemical
Engineering Research and Design 83 (6): 662–673.
Meredith, J. 1998. “Building Operations Management Theory
through Case and Field Research.” Journal of Operations
Management 16 (4): 441–454.
Neely, A., M. Gregory, and K. Platts. 1995. “Performance
Measurement System Design: A Literature Review and
Research Agenda.” International Journal of Operations &
Production Management 15 (4): 80–116.
New, S. 2007. “Celebrating the Enigma: The Continuing Puzzle
of the Toyota Production System.” International Journal of
Production Research 45 (16): 3545–3554.
Nicholas, J. M. 1998. Competitive Manufacturing Management:
Continuous Improvement, Lean Production, CustomerFocused Quality. Boston, MA: Irwin/ McGraw-Hill.
Papadopoulou, T., and M. Ozbayrak. 2005. “Leanness:
Experiences from the Journey to Date.” International Journal
of Manufacturing Technology Management 16 (7): 784–807.
Platts, K. W. 1993. “A Process Approach to Researching
Manufacturing Strategy.” International Journal of Operations
& Production Management 13 (8): 4–17.
Rother, M., and J. Shook. 1998. Learning to See: Value Stream
Mapping to Create Value and Eliminate Muda. Brookline,
MA: The Lean Enterprise Institute.
Samuel, D. E. 2011. “Exploring UK Lean Diffusion in the Period
1988 to 2010.” PhD thesis., Cardiff University.
Schmenner, R. W., and M. L. Swink. 1998. “On Theory in
Operations Management.” Journal of Operations Management
17 (1): 97–113.
Schonberger, R. J. 1982. Japanese Manufacturing Techniques:
Nine Hidden Lessons in Simplicity. New York: The Free
Press.
Schonberger, R. J. 2007. “Japanese Production Management: An
Evolution – With Mixed Success.” Journal of Operations
Management 25 (2): 403–419.
Serrano, I., C. Ochoa, and R. De Castro. 2008. “Evaluation of
Value Stream Mapping in Manufacturing System Redesign.”
International Journal of Production Research 46 (16): 4409–
4430.
Seth, D., and V. Gupta. 2005. “Application of Value Stream
Mapping for Lean Operations and Cycle Time Reduction:
An Indian Case Study.” Production Planning & Control 16 (1):
44–59.
Shah, R., and P. Ward. 2007. “Defining and Developing Measures
of Lean Production.” Journal of Operations Management 25
(4): 785–805.
Silverman, D. 2000. Doing Qualitative Research. London: Sage.
Sousa, R., and C. Voss. 2008. “Contingency Research in
Operations Management Practices.” Journal of Operations
Management 26 (6): 697–713.
Tapping, D., and T. Fabrizio. 2001. Value Stream Management:
Eight Steps to Planning, Mapping, and Sustaining Lean
Improvements. Portland, OR: Productivity Press.
Taylor, F. W. 1911. Principles of Scientific Management.
London: Harper & Brothers.
Vinodh, S., S. V. Kumar, and K. E. K. Vimal. 2014.
“Implementing Lean Sigma in an Indian Rotary Switches
Manufacturing Organisation.” Production Planning & Control
25 (4): 288–302.
Voss, C., N. Tsikriktsis, and M. Frohlich. 2002. “Case Research
in Operations Management.” International Journal of
Operations & Production Management 22 (2): 195–219.
Womack, J., and D. T. Jones. 1996. Lean Thinking: Banish Waste
and Create Wealth in Your Corporation. New York: Simon &
Schuster.
Womack, J., D. T. Jones, and D. Roos. 1990. The Machine that
Changed the World. New York: Simon & Schuster.
Wong, W. P., J. Ignatius, and K. L. Soh. 2014. “What is the
Leanness Level of Your Organisation in Lean Transformation
Implementation? An Integrated Lean Index Using ANP
Approach.” Production Planning & Control 25 (4):
273–287.
Worley, J. M., and T. L. Doolan. 2006. “The Role of
Communication and Management Support in Lean
Manufacturing Implementation.” Management Decision 44
(2): 228–245.
Yin, R. K. 2003. Case Study Research: Design and Methods. 3rd
ed. London: Sage.